FEATURE EXTRACTION AND CLASSIFICATION FOR SPAM DETECTION USING SUPERVISED ML

Abstract

The objective of this research is to enhance the efficacy and precision of spam detection in digital communication systems by utilising supervised machine learning techniques for feature extraction and classification procedures. This paper examines large email and text message datasets for feature extraction using preprocessing techniques such as tokenization, stop-word deletion, stemming, and vectorization. A number of supervised machine learning techniques, including as Naïve Bayes, Support Vector Machine (SVM), Decision Tree, Random Forest, and Logistic Regression, are compared in order to distinguish between spam and real messages. To measure how well these models work, we use metrics like F1-scores, recall, accuracy, and precision. In order to enhance classification accuracy and decrease false positives, the Paper stresses the significance of effective feature extraction methods as TF-IDF and Bag-of-Words. The proposed method allows trustworthy and secure communication by use of an adaptive spam detection framework that adjusts to actual changes in spam trends.